Page 69 - The prevalence of the Val66Met polymorphism in musicians: Possible evidence for compensatory neuroplasticity from a pilot study

108 | Cerebral Cortex, 2017, Vol. 27, No. 1

Table 2 Results of 2 × 2 ANCOVA, with sex as a covariate, indicating the main effects and an interaction of the COMT genotype and age on
Japanese language test scores
Source of variation SS df MS F P Remarks
COMT 152.857 1, 235 152.857 3.937 0.048* MM + VM > VV
Age 155.763 1, 235 155.763 4.012 0.046* young < old
COMT x age 174.686 1, 235 174.686 4.500 0.035*
Sex 3.387 1, 235 3.387 0.087 0.768 (n.s.)

Notes: Asterisks indicate significant results (P < 0.05). SS, sum of squares; df, degrees of freedom; MS, mean squares; and F, variance ratio. MM, Met/Met; VM, Val/Met;
VV, Val/Val.

performance by the Met carriers compared with the Val homo-
zygotes. Regarding the age effect, the old group exhibited a bet-
ter performance compared with the young group. A sex effect
was not identified. Importantly, a significant interaction was
identified between the COMT genotype and age.

Post hoc Results: COMT Genotype and Age Effects on Language
Test Scores
Because of the significant interaction detected between the
COMT genotype and age, post hoc simple main effect analyses
were performed using unpaired t-tests. The statistically signifi-
cant and non-significant differences in the language test scores
between the Met carriers (MM + VM) and Val homozygotes (VV)
for the young and old groups, as well as those between the
young and old groups for the Met carriers (MM + VM) and Val
Figure 1. Effects of COMT genotype and age on language test scores. A language
homozygotes (VV), are summarized in Figure 1. A significant
test was administered to assess comprehensive language ability. A significant
main effect of the COMT genotype on the language test scores
main effect of the COMT genotype on the language test score was identified
was identified between the Met carriers and Val homozygotes between the Met carriers (MM + VM) and Val homozygotes (VV) (t(92.5) = 2.628,
(t (92.5) = 2.628, P = 0.010**, MM + VM > VV) in the young group. P = 0.010**, MM + VM > VV) in the young group, whereas no main effect of the Downloaded from https://academic.oup.com/cercor/article-abstract/27/1/104/2617708 by guest on 24 November 2018
By contrast, no main effect of the COMT genotype was identi- COMT genotype was identified in the old group. Regarding the age effect on the
fied in the old group. Additionally, while a significant age effect language test score, no main effect was identified between the young and old
groups for the Met carriers (MM + VM); however, a significant main effect of age
on the test scores was identified among the Val homozygotes
was identified between the young and old groups for the Val homozygotes (VV)
(t (94.8) = −2.686, P = 0.009**, young < old), no age effect was found
(t(94.8) = −2.686, P = 0.009**, young < old). ASDS, adjusted standard deviation
among the Met carriers. scores. Error bars indicate standard error (SE). Asterisks indicate significant
results, and n.s. indicates not significant.
Cortical Responses During Language Processing
COMT Genotype Effects on Cortical Activation During in the angular gyrus and temporal region, including Wernicke’s
Language Processing area. There were no significant main or interaction effects of
Differences in cortical activation between the Met carriers and the COMT genotype in the supramarginal gyrus or frontal
Val homozygotes during word processing were compared via region, including Broca’s area. Therefore, these 2 ROIs were not
ROI-based analyses (4 ROIs: TR, AG, SMG, and FR). As an included in further analyses. In addition to the main effects of
example, the temporal dynamics of cortical activation are illu- the COMT genotype, an interaction between the COMT genotype
strated in Figure 2. Both genotype groups exhibited increases in and the task condition was identified in the angular gyrus (P <
[oxy-Hb] and decreases in [deoxy-Hb], which represents a 0.05, details in Supplementary Table 2). Because of this inter-
response pattern consistent with previous studies (Sakatani action and the finding that the differences in cortical responses
et al. 1998). This typical time course of [oxy-Hb] and [deoxy-Hb] between task conditions in the temporal region were not small
of grand-averaged data demonstrates increased cortical activa- (although they were not significant), post hoc analyses were
tion in the temporal region of Met carriers compared with Val performed for 2 separate task conditions for the 2 ROIs in
homozygotes. which the genotype effects were identified. The factors age,
We initially conducted 4-way repeated-measures ANCOVAs hemisphere, and sex, which had no significant interactions
for 4 ROIs, with sex as a covariate, to examine the effects of the with the COMT genotype, were not included in the post hoc
COMT genotype (Met carriers (MM + VM) and Val homozygotes analyses. Hereafter, P values were FDR corrected for 2 tests (for
(VV)), age group (young and old), task condition (high-frequency 2 ROIs) with a significance level of P < 0.05 after multiple testing
and low-frequency word conditions), and hemisphere (left and correction. As summarized in Figure 3, the unpaired t-test
right hemispheres). P values were FDR corrected for 4 tests (for results indicated significant differences in the cortical activa-
4 ROIs) with a significance level of P < 0.05 after multiple testing tion between the Met carriers (MM + VM) and Val homozygotes
correction. Because of space limitations, the complete results (VV) in the angular gyrus (t(244) = 3.552, uncorrected P <
are presented in Supplementary Table 2, and only the results of 0.001***, FDR-corrected P < 0.001***, MM + VM > VV) and the
the COMT genotype effects in 4 ROIs are listed in Table 3. temporal region, including Wernicke’sarea (t(244) = 2.947, uncor-
Significant main effects of the COMT genotype were identified rected, P = 0.0035**, FDR-corrected P = 0.0035**, MM + VM > VV)